Influences of weather on pedestrian safety perception at mid-block crossing: A CAVE-based study.

Reckless crossing behaviour is one of the major contributing factors to pedestrian crashes and injuries. The relationship between perceived risk and actual behaviour of pedestrians was examined. However, influences of weather conditions, which is a significant crash contributory factor, on the pedestrian safety perception are less studied. In this study, pedestrian safety perception in adverse weather and low visibility conditions like rain and fog is examined using immersive Cave Automatic Virtual Environment (CAVE) experiment. For instance, the 3D virtual reality model of a mid-block crossing in Hong Kong is developed. Factors including pedestrian socio-demographics, vehicle speed, gap size and weather condition are considered in the experiments. The propensity score method is adopted to estimate the causal inferences of weather conditions on pedestrian safety perception. Moreover, effects of multilevel data for multiple treatments are accounted using inverse probability of treatment weighting. Results indicate that perceived risk of pedestrians are higher in rainy and foggy conditions. Also, adverse impact of rainy condition is more significant in the dusk time, compared to daytime. Findings should shed light on effective remedial measures like traffic management and control, and street lighting that can mitigate the risk of pedestrian crash at the mid-block crossing.